This is a Preprint and has not been peer reviewed. This is version 2 of this Preprint.
The Vulnerability Spiral: Compound Climate Risk Across Five Islands
Downloads
Supplementary Files
Authors
Abstract
sea level rise stacked on top of degrading natural coastal defenses. That assumption is rarely tested against
independent, multi temporal evidence, and rarely broken down by ecosystem type, so I tested it directly,
across 5 islands spanning 3 ocean basins (Maldives, Lakshadweep, Seychelles, Fiji, and the Canary Islands).
I treated mangrove decline and coral reef degradation as 2 separate pathways, always kept apart, never
folded into a single mixed category. Physical exposure came from settlement level elevation data;
mangrove extent I tracked across 3 independent time points (1996, 2010, 2020); coral condition came from
a 24 year, monthly sampled satellite derived thermal stress record (1996–2020). Coral reef systems show a
measurable, rising bleaching stress trend in 4 of 5 islands, most severely in Seychelles. Mangroves,
assessed at the same 3 independent time points but without a comparable continuous trend test, show no
measurable decline in any of the 3 islands where they're present, so the assumption that ecosystem decline
is uniform doesn't hold up. Combine physical exposure with the coral thermal stress trend into a composite
vulnerability score, and Seychelles, the highest exposure island on the settlement based measure in this
sample, stays the highest overall compound risk once coral thermal stress is factored in too. Population
weighted exposure, after fixing an incomplete boundary polygon that had been letting open water leak into
the calculation (see Section 6), now shows a different order. Maldives (14.9%) sits marginally ahead of
Seychelles (14.2%), close enough that I read it as a near tie between the 2 highest exposure islands and not
a clean disagreement with the other measures. An earlier version of this analysis, built on a DEM data
quality problem later traced and fixed, made Maldives and Lakshadweep look far more exposed than they
are, creating an appearance of disagreement between indicators that didn't actually reflect anything real
once the underlying data was fixed. I also ran a supplementary test of whether formal protected area
coverage aligns with this empirically verified risk, and found a positive relationship whose p value crosses
the conventional 0.05 threshold on this small, 4 island sample (r=0.965, p=0.035). It's a result I still read
with caution, since it rests on so few data points. It's a strong signal on its own, but it doesn't confirm that
governance actually responds to risk. A descriptive comparison of settlement expansion (2016–2024) adds
a 3rd corroborating signal, since Seychelles shows the clearest built up increase in its broader settled area
while Fiji shows none. In short, coastal ecosystem degradation doesn't happen the same way across
ecosystem types, and that has a direct effect on how adaptation and conservation resources should be
prioritized.
DOI
https://doi.org/10.31223/X5M50Q
Subjects
Biodiversity, Earth Sciences, Environmental Monitoring, Environmental Sciences, Environmental Studies, Geographic Information Sciences, Geography, Geomorphology, Hydrology, International and Area Studies, Marine Biology, Natural Resource Economics, Nature and Society Relations, Oceanography and Atmospheric Sciences and Meteorology, Other Environmental Sciences, Other Geography, Remote Sensing, Spatial Science, Statistical Methodology, Statistical Models, Sustainability
Keywords
SEA LEVEL RISE, COASTAL VULNERABILITY, CORAL REEF BLEACHING, DEGREE HEATING WEEK, SMALL ISLAND STATES, MANGROVE EXTENT, COMPOUND VULNERABILITY SCORE, PROTECTED AREA GOVERNANCE, GEOSPATIAL REMOTE SENSING, SETTLEMENT ENCROACHMENT
Dates
Published: 2026-09-05 08:41
Last Updated: 2026-09-11 06:02
Older Versions
License
CC BY Attribution 4.0 International
Additional Metadata
Conflict of interest statement:
None
Data Availability:
https://github.com/sakshimaske303-commits/DOUBLE-JEOPARDY
Metrics
Views: 259
Downloads: 59
There are no comments or no comments have been made public for this article.